Afadin is localized to the OS, OLM, OPL, INL, and IPL.

A. Immunostaining of the wild-type (WT) mouse retina (1M) with anti-afadin antibody. The afadin signals disappeared in the pre-absorbed (peptide+) sample. The signal observed in the peptide+ image represents the background and non-specific staining. Nuclei were stained with DAPI (blue). B. Immunostaining of the WT retinal section (1M) using anti-afadin (green) combined with anti-Nectin-1 (red) antibody. Necin-1 was partially co-localized with afadin in the OLM, OPL, and IPL. C. Afadin was colocalized with nectin-1, nectin-2, and nectin-3 at the OLM. The 1M WT retinal sections were immunostained with anti-afadin (green), anti-nectin-1 (red, upper panels), anti-nectin-2 (red, middle panels), and anti-nectin-3 (red, lower panels) antibodies. D. Immunostaining of 1M WT retinal sections using anti-afadin (green), anti-SCGN (red), and anti-PKCα(white) antibodies. Afadin signals overlapped with SCGN and PKCα in the OPL and INL, and PKCα in the IPL. OS, the outer segment; IS, the inner segment; OLM, the outer limiting membrane; ONL, the outer nuclear layer; OPL, the outer plexiform layer; INL, the inner nuclear layer; IPL, the inner plexiform layer; GCL, the ganglion cell layer.

Outer retinal lamination is severely disrupted in the afadin cKO retina.

A. Toluidine blue staining of the cHet and cKO mouse retinal sections at 1M. The retinal layer structure was disrupted in the cKO retina. Rod photoreceptor nuclei (insets) were deeply stained. The red asterisk indicates a rod photoreceptor. The arrowhead indicates a rosette-like structure. Scale bar in the inset, 2.5 μm. B. Representative images of the cHet and cKO retinas (1M) stained with anti-Rom1 (green) and anti-Rhodopsin (Rho, red) antibodies and PNA-rhodamine (white). Rom1, Rhodopsin, and PNA signals were remarkably decreased and scattered in the cKO mice. The arrowhead indicates the rosette-like structure. C. Representative images of the cHet and cKO retinas (1M) stained with anti-S-opsin (green) and anti-M-opsin (white) antibodies and PNA-rhodamine (red). Cone OSs were aberrantly located in the cKO retina. The arrowhead indicates the rosette-like structure. D. Retinal flat-mount immunostaining of the cHet and cKO retinas immunostained with anti-Arr3 (green, left panels), anti-Rom1 (green, right panels), and anti-Rhodopsin (red, right panels) antibodies. The inset, a magnified view of the area enclosed by the dashed box, shows an area with a relatively high density of photoreceptors in the cKO retina. E. Immunofluorescent analysis of the cHet and cKO retinas (1M) using anti-Arr3 (cone marker, green in the left panels), anti-Otx2 (photoreceptor and BC marker, red in the left panels), anti-RNA binding protein with multiple splicing (RBPMS, RGC marker, green in second left panels), anti-Chx10 (BC marker, red in second left panels), anti-Lhx2 (Müller glia marker, green in third left and third right panels), anti-AP2α (AC marker, red in third left and second right panels), and anti-Calbindin (horizontal cell and AC marker, white in third left and right panels) antibodies. Arrowheads indicate RGCs and ACs near the rosette-like structure. F-L. Distribution of each retinal cell type based on their distance from the retinal surface in the cHet and cKO mice. The shortest distance between each cell body and either the outer or inner retinal surface was measured using ImageJ. For cones (F), rods (G), bipolar cells (H), horizontal cells (I), and Müller glia (J), the distance was measured from the outer retinal surface. For amacrine cells (K) and retinal ganglion cells (L), the distance was measured from the inner retinal surface. Note: In K and L, the X-axis is reversed such that short distance appears on the right. The outer retinal surface was defined as the region immediately above the somata of the outer retinal cells. Bin width was 5 μm (0– 5 up to 175–180 μm). Inset box plots show the distribution of raw values, with whiskers extending to 1.5 times the interquartile range and dots indicating outliers (cHet; Cone 3.45 ± 2.72 μm, n = 141 cells from 4 mice, Rod 23.1 ± 13.8 μm, n = 3055 cells from 4 mice, Bipolar cell 59.4 ± 6.74 μm, n = 432 cells from 3 mice, Horizontal cell 57.5 ± 6.45 μm, n = 20 cells from 3 mice, Müller glial cell 69.9 ± 7.61 μm, n = 191 cells from 3 mice, Amacrine cell 51.8 ± 18.6 μm, n = 362 cells from 3 mice, Ganglion cell 8.71 ± 3.71 μm, n = 93 cells from 3 mice, cKO; Cone 14.9 ± 14.2 μm, n = 42 cells from 4 mice, Rod 35.5 ± 30.5 μm, n = 716 cells from 4 mice, Bipolar cell 56.6 ± 32.7 μm, n = 1236 cells from 5 mice, Horizontal cell 37.2 ± 31.2 μm, n = 14 cells from 4 mice, Müller glial cell 45.9 ± 22.9 μm, n = 95 cells from 4 mice, Amacrine cell 63.8 ± 34.6 μm, n = 265 cells from 4 mice, Ganglion cell 38.9 ± 33.6 μm, n = 116 cells from 5 mice). ***p < 0.001, **p < 0.01 by a generalized linear mixed model (GLMM) or log1p-transformation followed by a linear mixed model (LMM). Statistical tests: Cone, Rod, log1p-transformation followed by a LMM; Bipolar cell, Horizontal cell, Müller glial cell, Amacrine cell, Ganglion cell, GLMM. M. The thickness of the cHet and cKO retinas at 1M. (cHet; 137.9 ± 11.1 μm, n = 4, cKO; 135.5 ± 9.1 μm, n = 5). Error bars, mean ± SD. No significant difference was observed between the two groups (p = 0.735 by Student’s t-test).

Photoreceptor-BC synapses were severely impaired and mislocalized in the afadin cKO retina.

A. Immunostaining of the cHet and cKO retinal sections at 1M with anti-Bassoon (green), anti-mGluR6 (red), and anti-PKCα (white) antibodies. The arrowhead indicates the synapse between rod and rod-BC. Small panels are magnified views of the area enclosed by the dashed box. B. Immunostaining of the cHet and cKO retinas at 1M with anti-GluR5 (green), anti-PSD95 (red), and anti-PKARIIβ (white) antibodies. The arrowhead indicates the synapse between the cone and OFF-BC. Small panels are magnified views of the area enclosed by the dashed box. C. 3D projection of confocal image immunostaining with anti-Bassoon (green), anti-mGluR6 (red), and anti-PKCα (white) using the cHet and cKO retina (1M). The inset, a magnified view of the area enclosed by the dashed box, shows the synapse between the rod and BC at high magnification. The arrowhead indicates the ribbon synapse between rod and rod-BC. D. 3D projection of confocal image immunostaining with anti-GluR5 (green), anti-PSD95 (red), and anti-PKARIIβ (white) antibodies using the cHet and cKO retina (1M). E. Quantification of the number of synapses between rod photoreceptor and rod-BC under 1mm2 of retinal surface in the cHet and cKO mice immunostained with anti-Bassoon, anti-mGluR6, and anti-PKCα antibodies (cHet; 599.6 ± 49.7, n = 3, cKO; 29.1 ± 12.75, n = 4, 10 images from each mouse. Error bars, mean ± SD. ***p < 0.001 by a GLMM). The number of synapses decreased to about 5 % of the cHet in the cKO retina. F. Quantification of the number of synapses between photoreceptor cells and type 3 OFF-BC under 1mm2 of the retinal surface in the cHet and cKO mice immunostained with anti-GluR5, anti-PSD95, and anti-PKARIIβ antibodies (cHet; 24.0 ± 2.8, n = 3, cKO; 4.1 ± 1.3, n = 3, 10 images from each mouse. Error bars, mean ± SD. ***p < 0.001 by a GLMM). The number of synapses between the photoreceptor and Type 3 OFF-BC in the cKO retina decreased to about 15% of the cHet retina.

GluR5 is ectopically localized to ON and OFF BC processes.

A. Immunostaining of the cHet and cKO retinal sections at 1M with anti-GluR5 (green), anti-SCGN (red), and anti-PKCα antibodies. GluR5 signal was obviously observed in the IPL of the cKO retina. Yellow and white arrowheads indicate overlap of GluR5 and PKCα, and GluR5 and SCGN, respectively. Small panels are magnified views of the area enclosed by the dashed box. B. Representative images of the immunostained cHet and cKO retinas (1M) with anti-GluR5 (green), anti-vGlut1 (red), and anti-SCGN (white) antibodies. C. Immunofluorescent analysis of the cHet and cKO retinas (1M) with anti-HPC-1 (AC marker, green in the left panels), anti-Calbindin (red in the left panels), anti-Tuj1 (RGC marker, green in the right panels), and anti-RBPMS (red in the right panels) antibodies. Arrowhead indicates acellular patches. D. The number of each retinal cell type per 100 μm width of retinal section at 1M (cHet; Rod 185.9 ± 17.5, Cone 8.4 ± 1.4, Bipolar cell 45.5 ± 5.2, Ganglion cell 7.4 ± 1.5, Horizontal cell 2.1 ± 0.4, Amacrine cell 30.1 ± 3.2, Müller glial cell 15.7 ± 2.3, n = 4, cKO; Rod 57.3 ± 17.3, Cone 3.5 ± 1.3, Bipolar cell 64.6 ± 7.8, Horizontal cell 1.9 ± 0.7, Amacrine cell 27.6 ± 6.3, Ganglion cell 5.7 ± 0.9, Müller glial cell 12.4 ± 1.5, n = 4. Error bars, mean ± SD. Horizontal cell p = 0.58, AC p = 0.52, RGC p =0.10. ***p < 0.001, **p < 0.01, *p < 0.05 by Student’s t test. In the cKO retina, rod and cone photoreceptors and Müller glial cells significantly decreased to about 30, 40, and 80% of the cHet retina, respectively, and BCs significantly increased to about 150% of the cHet retina. E. Immunofluorescent analysis of the cHet and cKO retinas at P1, P3, P5, P8, P11, and P14 using anti-Chx10 (green), anti-Otx2 (red), and anti-Ki67 (progenitor cell marker, white) antibodies. NBL, the neuroblastic layer.

Reduction of the a- and b-waves in the cKO retina.

A. Left, Scotopic ERGs from the anesthetized cHet (n = 3; black) and cKO (n = 3; magenta) mice. A flashlight (duration; 5 ms, white LED, intensity; 1.0 × 104 cd/m2) was applied to eyes in the dark three times. Traces: averaged (dark) and SD (pale). Right, Photopic ERGs. The anesthetized cHet (black) and cKO (magenta) mice were light (31.6 cd/m2) adapted for 10 min, and then, a flashlight (duration; 5 ms, intensity; 1.0 × 104 cd/m2) was superimposed on the adapted light to eyes 16 times. B. mERGs recorded by MEA from the cHet (black) and cKO (magenta) retinas. A flashlight (duration; 5 ms, green LED; λmax = 510 nm, intensity; 2.5 × 104 photon/s/µm2) was applied to the isolated whole retina 3 times every 10 s. Control (solid line), L-AP4 (10 μM) (pale), and after washout (dotted line). Left, cHet; SD of the amplitude for 200 ms after the flash onset; >1,000 μV, 292/413 electrodes, n = 7 retinas from 4 mice. Right, cKO; SD of the amplitude for 300 ms after the flash onset; >50 μV, 13/590 electrodes, n = 10 retinas from 6 mice. C. Left, Amplitude of the a-wave (cHet; 155 ± 58.7 μV, n = 292 electrodes, n = 7 retinas from 4 mice, cKO; 44.1 ± 21.7 μV, n = 13 electrodes, n = 10 retinas from 6 mice, Error bars, mean ± SD. ***p < 0.001, Mann-Whitney U test). Right, Amplitude of the b-wave (cHet; 163 ± 76.5 μV, n = 292 electrodes, cKO; 28.4 ± 19.6 μV, n = 13 electrodes, p = 7.08 x 10-9, Mann-Whitney U test). D. Ratio of b-wave amplitude / a-wave amplitude (b / a) (cHet; 1.19 ± 0.67, cKO; 0.59 ± 0.23, Error bars, mean ± SD. ***p < 0.001, Mann-Whitney U test). E. Left, Implicit time of the a-wave (cHet; 62.6 ± 10.8 ms, n = 292 electrodes, cKO; 149 ± 16.0 ms, n = 13 electrodes, p = 5.34 x 10-10, Mann-Whitney U test). Right, Implicit time of the b-wave (cHet; 127 ± 25.3 ms, n = 292 electrodes, cKO; 229 ± 60.7 ms, n = 13 electrodes, Error bars, mean ± SD. ***p < 0.001, Mann-Whitney U test). F. Difference between implicit time of the b-wave and that of the a-wave (cHet; 64.4 ± 22.0 ms, n = 292 electrodes, cKO; 80.4 ± 50.9 ms, n = 13 electrodes, Error bars, mean ± SD. p = 0.40, Mann-Whitney U test).

RGC classification based on the light-evoked responses.

A,B. Representative light-evoked responses. A, The cHet retina. B, The cKO retina. A flash (duration; 2 s, intensity; 24.2 or 27.3 cd/m2) was applied seven times every 8 s. After spike sorting, the raster plots (top) and the PSTHs (bottom) were created (20 ms/bin). Based on the PSTHs, RGC responses were classified into “ON”, “ON-OFF”, “ON-OFF inhibition”, “OFF”, and “None” types. C. Ratio of responded to non-responded types (cHet; n = 434, cKO; n = 388). D. Rod and cone inputs to RGCs (cHet; n = 434, cKO; n = 388). 2-s green LED flash (λmax = 510 nm, intensity; 4.8 x 103 photons/s/µm2) and UV LED flash (λmax = 360 nm, intensity; 3.3 x 104 photons/s/µm2) were applied 7 times every 8 s to stimulate mainly rods and S cones, respectively.

Spatiotemporal properties of the RF of RGCs.

A. Left, An isolated retina on the MEA. Electrodes (black square), mapped RFs (colored oval), and the region where immunohistochemical examination was performed after recording (magenta square). Scale bar 200 μm. Right, Immunohistochemical staining with anti-PKARIIβ (green), anti-PSD95 (red), and anti-PKCα (white) antibodies. IPL: the inner plexiform layer. B. Number of cells whose RF was clearly observed (cHet; 195/434 cells, n = 4 retinas, cKO; 69/388 cells, n = 4 retinas). C. The spatial profile of RFs in the cHet (Top) and cKO (Bottom) retinas. Color scale illustrates high (red) to low (blue) relative to the average (green). Scale bar 100 μm. D. RF size of the cKO retina was significantly smaller than that of the cHet retina (cHet; 0.0307 ± 0.0210 mm2 n = 195, n = 4 retinas, cKO; 0.0161 ± 0.00839 mm2 n = 69, n = 4 retinas, Error bars, mean ± SD. ***p < 0.001, Mann-Whitney U test). E. Short and long axes of the oval fitted to each RF. F. Clustering of RGCs based on the temporal profile of RFs. Five clusters were visualized (cHet; circle, n = 4 retinas, cKO; x, n = 4 retinas). G. Temporal profiles from each cluster were superimposed separately (cHet; black, cKO; magenta, mean; solid line, SD: pale line). H. Time-to-peak latency of the temporal profile (cHet; filled bar, cKO; open bar). Mean ± SD. (cHet; cluster 1; -113 ± 15 msec, n = 31, cluster 2; -111 ± 13 msec, n = 36, cluster 3; -200 ± 31 msec, n = 51, cluster 4; -171 ± 19 msec, n = 37, cluster 5; -178 ± 28 msec, n = 40, cKO; cluster 3; -163 ± 21 msec, n = 10, cluster 4; -238 ± 34 msec, n = 36, cluster 5; -217 ± 41 msec, n = 23).

Visual function is partially retained in the afadin cKO mice.

A. Schematic diagram of the experimental setup for the OMR test. Drifting square-wave gratings were displayed on monitors surrounding the mouse, and head-tracking movements were recorded under photopic conditions. B. Representative traces of head-tracking movements in the cHet, cKO, and ONC mice. Blue lines indicate head movement trajectories, and red segments denote the periods when the head movement aligned with the direction and timing of the visual stimulus. C. Summary of head-tracking movements in response to drifting gratings with spatial frequencies of 0.056, 0.10, and 0.27 cycles/degree, as well as under no-stimulus conditions, in the cHet, cKO, and ONC mice (cHet; 0.056 cycles/degree 7.30 ± 1.94, 0.10 cycles/degree 7.89 ± 2.00, 0.27 cycles/degree 7.16 ± 1.39, n = 8 mice, cKO; 0.056 cycles/degree 2.36 ± 1.59, 0.10 cycles/degree 1.64 ± 1.29, 0.27 cycles/degree 0.0268 ± 0.103, n = 14 mice, ONC; 0.056 cycles/degree 0.0833 ± 0.121, 0.10 cycles/degree 0.0714 ± 0.161, 0.27 cycles/degree 0.0476 ± 0.128, n = 7 mice). Error bars, mean ± SD. No stimulus; cHet vs cKO p > 0.9999, cHet vs ONC p >0.9999, cKO vs ONC p > 0.9999, 0.27 cycles/degree; cKO vs ONC p > 0.9999. ***p < 0.001, Dunn’s test. D. Comparison of head-tracking responses across spatial frequencies (0.056, 0.10, and 0.27 cycles/degree). Statistical comparisons were conducted within each mouse group. No significant differences were detected across spatial frequencies in cHet and ONC mice. In contrast, the cKO mice exhibited a trend toward decreased head-tracking responses with increasing spatial frequency. Error bars, mean ± SD. cHet; 0.056 cycles/degree vs 0.1 cycles/degree p = 0.3566, 0.1 cycles/degree vs 0.27 cycles/degree p = 0.2606, 0.056 cycles/degree vs 0.27 cycles/degree p > 0.9999, cKO; 0.056 cycles/degree vs 0.1 cycles/degree p = 0.2577. ***p < 0.001, Dunn’s test. E. Schematic diagram of the visual cliff test. The apparatus consists of a single transparent platform spanning two arenas with identical checkerboard patterns: one immediately beneath the platform (the shallow side) and the other on the floor (the deep side), creating a visual cliff. Mice are placed at the boundary between the two sides, and their location preference is evaluated. F. Time spent on the shallow side during the visual cliff test. The percentage of time spent on the shallow side was significantly higher in the cHet mice than in the cKO and ONC mice. No significant difference was observed between the cKO and ONC mice (cHet; 81.1 ± 14.6, n = 19 mice, cKO; 57.4 ± 17.6, n = 26 mice, ONC; 51.1 ± 6.38, n = 6 mice). Error bars, mean ± SD. ***p < 0.001, one-way ANOVA followed by Tukey’s multiple comparisons test.

Ectopic AJs are observed in the developing afadin cKO retina.

A. Immunostaining of the cHet and cKO retinas at postnatal day 0 (P0) with anti-phospho-histone H3S10 (pHH3, mitotic cell marker, green), anti-N-Cadherin (red in left, second left, and second right panels), anti-Nectin-1 (white in left, second left, and third right panels), anti-β-catenin (red in the right panels) antibodies. Ectopic N-cadherin, nectin-1, and β-catenin signals were observed in the cKO retinas, and pHH3-positive cells were localized near these ectopic signals. B. Quantification of pHH3-positive cells per 100 μm width of retinal sections in the cHet and cKO retinas at P0 (cHet; 4.40 ± 0.937, n = 3, cKO; 6.69 ± 0.507, n = 4). pHH3 positive cells significantly increased in the cKO retina. Error bars, mean ± SD. **p < 0.01 by Student’s t-test). C. Retinal sections from the cHet and cKO mice immunostained with anti-pHH3 antibody (red) at P1 (upper panels), P3 (middle panels), and P5 (lower panels). D. Distribution of mitotic cells based on their distance from the outer retinal surface in the cHet and cKO mice at P1 (top), P3 (middle), and P5 (bottom). The shortest distance between mitotic cell body and the outer retinal surface was measured using ImageJ. Bin width was 5 μm (0–5 up to 245–250 μm). Inset box plots show the distribution of raw values, with whiskers extending to 1.5 times the interquartile range and dots indicating outliers (cHet; P1 3.76 ± 17.5 μm, n = 300 cells from 5 mice, P3 6.73 ± 28.8 μm, n = 469 cells from 3 mice, P5 16.7 ± 44.7 μm, n = 180 cells from 4 mice, cKO; P1 48.4 ± 24.1 μm, n = 326 cells from 4 mice, P3 76.0 ± 41.9 μm, n = 902 cells from 4 mice, P5 94.0 ± 44.8 μm, n = 163 cells from 4 mice). ***p < 0.001 by log1p-transformation followed by a LMM. E. Retinal sections from the cHet and cKO mice immunostained with anti-BrdU antibody (green) at P1 (upper panels), P3 (middle panels), and P5 (lower panels). F. Distribution of progenitor cells based on their distance from the outer retinal surface in the cHet and cKO mice at P1 (top), P3 (middle), and P5 (bottom). Bin width was 5 μm (0–5 up to 245–250 μm). Inset box plots show the distribution of raw values, with whiskers extending to 1.5 times the interquartile range and dots indicating outliers (cHet; P1 58.0 ± 22.1 μm, n = 2527 cells from 4 mice, P3 83.5 ± 31.7 μm, n = 1904 cells from 4 mice, P5 95.3 ± 32.4 μm, n = 620 cells from 5 mice, cKO; P1 62.7 ± 35.2 μm, n = 1713 cells from 4 mice, P3 96.5 ± 48.5 μm, n = 2025 cells from 3 mice, P5 90.7 ± 40.5 μm, n = 630 cells from 3 mice). P5 p = 0.49, ***p < 0.001, **p < 0.01 by log1p-transformation followed by a LMM.

Afadin is localized to AJs in developing and mature retinas.

A. Immunostaining of the WT retinal section (1 M) with anti-afadin (green) antibody combined with nectin-2 (red, upper and middle panels) and nectin-3 (red, lower panels) antibodies. Nectin-2 and nectin-3 were localized in the OLM. The nectin-2 signal in the IPL was insufficient for reliable assessment of its localization and colocalization. B. Immunostaining with anti-afadin (green), anti-nectin-1 (red), and anti-Arr3 (a cone marker, white) antibodies. Afadin was co-localized with nectin-1 in cone synapses. C. Immunostaining of the WT retinal section (P0) with anti-afadin (green) and anti-nectin-1 (red) antibodies.

AJs in the OLM are disrupted in the afadin cKO retina.

A, B. Immunostaining of Dkk3-Cre; R26R-H2B-EGFP and R26R-H2B-EGFP retinas at P0 (A) and 1M (B) using an anti-GFP antibody (green). The Cre recombinase activity was detected in the Dkk3-Cre; R26R-H2B-EGFP retina at both stages. C. Western blot analysis of the cHet and cKO retinas using anti-afadin (upper panel) and anti-GAPDH (lower panel) antibodies. No significant afadin band was detected in the cKO retina. D. Electron microscopic analysis of the cHet and cKO retinas (1M). A few ectopic disc structures were observed in the cKO retina. The inset, a magnified view of the area enclosed by the dashed box, shows the ectopic disc structure at high magnification. The arrowhead indicates the outer segment disc structure. RPE, the retinal pigment epithelium. E. Immunostaining of the cHet and cKO retinal sections at 1M using anti-β-catenin (green, left panels), anti-nectin-1 (red, left panels), and anti-N-cadherin (green, right panels). Signals of these markers were not observed at the outer retinal surface in the cKO mice. F. Immunostaining of the cHet and cKO retinal sections at 1M using anti-glutamine synthetase (GS, green) antibody. Obvious GS signals were observed in the inner side of the retina but not at the outer retinal surface in the cKO mice. G. Retinal sections from the cHet and cKO mice at 1M were stained with an anti-CD31 antibody (red). No notable vascular differences were observed in the cKO retina.

The OS and photoreceptor-BC synapses are affected in the developing afadin cKO retina.

A. Immunostaining of the cHet and cKO retinas at 1M with anti-PKCα (green), anti-PSD95 (red) antibodies. Small panels are magnified views of the area enclosed by the dashed box. B. 3D projection of confocal image immunostaining with anti-PKCα (green), anti-PSD95 (red) antibodies. C. Quantification of the number of contacts between photoreceptor and rod ON-BC under 1mm2 of retinal surface in the cHet and cKO mice immunostained with anti-GluR5, anti-PSD95, and anti-PKARIIβ antibodies (cHet; 550.5 ± 29.1, n = 3, cKO; 65.2 ± 7.9, n = 3, 10 images from each mouse. Error bars, mean ± SD ***p < 0.001 by a LMM). The number of synapses between the rod and ON BC decreased to about 10% of the cHet in the cKO retina. D, E. The cHet and cKO retinas (1M) stained with anti-EAAT5 (green), anti-SCGN (red in D), and anti-PKCα (white in D), anti-PSD95 (red in E) antibodies and PNA-rhodamine (white in E). Glutamate transporter EAAT5 localization to ON and OFF BC axon terminals and photoreceptor axon terminals were unaffected in the cKO retina. Small panels are magnified views of the area enclosed by the dashed box. F. Immunostaining of the cHet and cKO retinal sections at 1M with anti-ChAT (green), anti-Calretinin (red), and anti-Calbindin (white) antibodies. Obvious ChAT bands, Calretinin bands, and Calbindin bands were observed in the IPL of the cKO mice.

Retinal cell fate determination is altered during development in the afadin cKO mice.

A. The number of PKCα+ and SCGN+ cells per 100 μm width of retinal sections at 1M (cHet; PKCα+ 15.8 ± 2.6, SCGN+ 24.7 ± 4.8, n = 4, cKO; PKCα+ 23.6 ± 5.2, SCGN+ 37.4 ± 7.7, n = 4. Error bars, mean ± SD. *p < 0.05 by Student’s t-test). PKCα+ cells and SCGN+ cells significantly increased to 150% of the cHet. B. The number of PKCα+ cell processes in the IPL per 100 μm width of retinal sections at 1M (cHet; 18.6 ± 0.5, n = 4, cKO; 14.7 ± 1.2, n = 4. Error bars, mean ± SD. ***p < 0.01 by Student’s t-test). The number of PKCα+ cell processes significantly decreased to 80% of the cHet. C. The number of each retinal cell types per 100 μm width of retinal sections at P14 (cHet; Rod 176.5± 14.3, Cone 10.6 ± 1.3, Bipolar cell 49.2 ± 1.9, Ganglion cell 8.1 ± 0.8, Horizontal cell 2.5 ± 0.7, Amacrine cell 30.3 ± 2.1, Müller glial cell 17.3 ± 1.4, n = 4, cKO; Rod 98.3 ± 3.8, Cone 4.7 ± 0.9, Bipolar cell 76.8 ± 10.3, Horizontal cell 1.9 ± 0.8, Amacrine cell 32.2 ± 8.9, Ganglion cell 8.5 ± 1.0, Müller glial cell 17.2 ± 3.3, n = 3. Error bars, mean ± SD. Horizontal cell p = 0.36, Amacrine cell p = 0.68, Ganglion cell p =0.57, Müller glial cell p = 0.94. ***p < 0.001 by Student’s t-test). Rod and cone photoreceptors were significantly reduced to approximately 55% and 45%, respectively, of those in the cHet retina, while BCs significantly increased to approximately 150% in the cKO retina. D. Expression of photoreceptor and BC marker genes in the cHet and cKO at P14 (Opn1sw and Opn1mw; cone maker, Nrl and Rhodopsin; rod marker, Chx10; BC marker, Trpm1 and mGluR6; ON BC marker). Each gene expression level was normalized to GAPDH, a housekeeping gene. The expression of Nrl, Rhodopsin, Opn1sw, and Opn1mw significantly decreased, whereas the expression of mGluR6, Trpm1, and Chx10 significantly increased in the cKO retina. Error bars, mean ± SD. ***p < 0.001, **p < 0.01. Statistical tests: Opn1sw Welch’s t-test; Opn1mw Student’s t-test; Nrl Welch’s t-test; Rho Student’s t-test; Chx10 Welch’s t-test; Trpm1 Student’s t-test; mGluR6 Student’s t-test. E. The number of BCs per 100 μm width of retinal sections at P1, P3, and P5 (cHet; P1 4.42 ± 1.09, n = 5, P3 17.2 ± 2.47, n = 3, P5 59.9 ± 7.42, n = 4, cKO; P1 4.71 ± 1.75, n = 4, P3 18.7 ± 1.74, n = 4, P5 96.4 ± 6.91, n = 3). BCs (Chx10-positive and Ki67-negative) were significantly increased in the cKO retina at P5, but not at P1 or P3. Error bars, mean ± SD. P1 p = 0.77, P3 p = 0.56, **p < 0.01 by Student’s t-test. F. Expression of rod photoreceptor cell marker genes (Rho and Nrl) in the cHet and cKO at P1, P3, and P5. Each gene expression was normalized to GAPDH, a housekeeping gene. The expression of Nrl and Rhodopsin significantly decreased in the cKO retina at P5. No significant difference was observed at P1 and P3. Error bars, mean ± SD. P1: Nrl p = 0.50, Rho p = 0.81, P3: Nrl p = 0.97, Rho p = 0.81, *p < 0.05. Statistical tests: P1 Rho, Mann–Whitney U test; P1 Nrl, P3 Rho, P3 Nrl, P5 Nrl, and P5 Rho, Student’s t-test.

Retinal lamination is disputed in the afadin cKO retina during development.

A. Immunostaining of the cHet and cKO retinal sections at P1, P3, P5, P8, P11, and P14 with anti-HPC-1 (green), anti-RBPMS (red) antibodies. Yellow and white arrowheads indicate acellular patches and aberrant acellular regions not containing IPL components, respectively. B. Immunostaining of the cHet and cKO retinal sections at P1, P3, P5, P8, P11, and P14 with anti-AP2α (green), anti-Calbindin (red) antibodies. Yellow and white arrowheads indicate acellular patches and aberrant acellular regions not containing IPL components, respectively. C. Immunostaining of the cHet and cKO retinal sections at P8, P11, and P14 with anti-Rom1 (green), anti-Rho (red) antibodies. Yellow and white arrowheads indicate the rosette-like structure and aberrant acellular regions not containing Rom1 and Rho signals, respectively. D. Representative images of the cHet and cKO retinas (P11) stained with antibodies against PKARIIβ (green, left panels), Arr3 (red, left panels), Bassoon (green, right panels), mGluR6 (red, right panels), PKCα (white, right panels). The inset, a magnified view of the area enclosed by the dashed box, shows the synapse between Cone and Type 3 OFF-BC (left panels), and ribbon synapses (right panels) at high magnification.

The number of apoptotic cells decreased in the developing cKO retina.

A. Immunostaining of the cHet and cKO retinas (1M) with anti-active caspase 3 (AC3, green) antibody. Arrowhead indicates AC3 positive cell. B. Quantification of apoptotic cells per 1mm2 in the cHet and cKO retinas at P1, P3, P5, P14, and 1M (cHet; P1 176.9 ± 31.0, n = 4, P3 164.7 ± 33.4, n = 3, P5 112.1 ± 25.8, n = 4, P14 13.7 ± 4.5, n = 5, 1M 0.577 ± 0.699, n = 4, cKO; P1 322.1 ± 65.5, n = 4, P3 cKO; 267.4 ± 29.1, n = 3, P5 177.1 ± 36.0, n = 4, P14 86.4 ± 17.3, n = 3, 1M 66.7 ± 8.2, n = 5). Apoptotic cells significantly increased in the cKO retina. Error bars, mean ± SD. *p < 0.05, **p < 0.01. Statistical tests: P1, P3, P5 Student’s t-test; P14 Mann–Whitney U test; 1M Welch’s t-test. C. Quantification of progenitor cells labeled by BrdU per 100 μm width of retinal section in the cHet and cKO retinas at P1, P3, and P5 (cHet; P1 70.8 ± 7.64, n = 4, P3 66.6 ± 17.0, n = 4, P5 17.6 ± 8.87, n = 3. cKO; P1 63.2 ± 7.31, n = 4, P3 65.1 ± 2.15, n = 3, P5 31.8 ± 5.03, n = 3). There was no significant difference between the cHet and cKO retinas. Error bars, mean ± SD. P1 p = 0.20, P3 p = 0.87, P5 p = 0.074. Statistical tests: P1, P5 Student’s t-test; P3 Welch’s t-test. D. Quantification of the retinal thickness in the cHet and cKO mice at P1, P3, and P5 (cHet; P1 145.1 ± 16.6, n = 5, P3 190.2 ± 9.81, n = 4, P5 203.9 ± 26.6, n = 5. cKO; P1 139.2 ± 26.1, n = 4, P3 203.7 ± 21.6, n = 5, P5 200.0 ± 26.7, n = 4). There was no significant difference between the cHet and cKO retinas. Error bars, mean ± SD. P1 p = 0.54, P3 p = 0.29, P5 p = 1.0. Statistical tests: P1, P3 Student’s t-test; P5 Mann–Whitney U test.